支持向量机在基因表达数据分类中的应用研究  

The Research of SVM Applied to the Classification of Gene Expression Data

作　　者：武振宇[1] 李康[1] 

机构地区：[1]哈尔滨医科大学卫生统计学教研室,研究生150086

出　　处：《中国卫生统计》2007年第1期8-11,共4页Chinese Journal of Health Statistics

基　　金：国家自然科学基金资助(30371253);黑龙江省重点项目(GB04C30202)

摘　　要：目的探讨支持向量机在基因表达数据分类研究中的应用条件和效果。方法使用支持向量机软件包,通过实际基因表达数据考核其应用效果,并通过模拟试验进一步验证和研究在含有大量无差异表达基因情况下对分类产生的影响。结果对四种疾病的真实基因表达数据的分类取得了良好的效果,模拟试验则显示了支持向量机对分类具有较高的准确性,但随无差异基因数量的增加其分类效果呈明显下降的趋势;在类间分离一定的情况下,差异表达基因数目较多、基因之间具有较高的相关性时,更容易获得好的分类效果。结论支持向量机在解决小样本、非线性及高维问题中表现出许多潜在的优势,可以有效地用于分析基因表达数据的分类问题。Discuss the condition and the effect of SVM in the classification of gene expression data. Methods Using thepackage of SVM to test the effect of classification according to the real gene expression data. Simulation studies are conducted to validate and investigate the influence of the huge number of noises on the classification. Results SVM achieves good outcomes in the classification of four real gene expression data. The simulation tests show that the predicted accuracy performed by SVM decreases while the number of noise increases. SVM performs especially well when the differences between two groups are determinate, the number of differential gene is high, and the differential genes are strongly related. Conclusion There are lots of potential advantages of SVM in solving the problem of classification in the nonlinear, high dimension, limited-case samples, and SVM could be effectively used in the field of classification of gene expression data.

关 键 词：支持向量机 基因表达数据 分类研究 模拟试验 

分 类 号：Q78[生物学—分子生物学]